Physiological pharmacokinetic models are developed for the distribution and disposition of drugs and environmental contaminants. These models provide a plausible set of equations that can be used to extrapolate data from experimental systems to humans and provide a rational basis for improvement of chemotherapy, chemoprevention, and risk assessment. Regional delivery has been emphasized with specific work on delivery of drugs to the central nervous system, peritoneal cavity and lung. (1) A protein toxin has been shown to have significant activity in a phase I clinical trial of recurrent glioblastoma when the agent is administered by direct infusion. A major problem has been the inability to track the spatial and temporal distribution of the infused protein. Work in progress has examined the applicability of imaging methods to follow the distribution of infused material. Images from single-photon emission tomography, computerized tomography and magnetic resonance have been compared with the distribution of proteins determined by quantitative autoradiography in the monkey brain. (2) Peritoneal drug administration can provide a large pharmacokinetic advantage. The pharmacokinetic theory has been validated in numerous clinical studies, and a significant survival advantage has been demonstrated in a published prospective, randomized phase III clinical trial in which the only variable was the route of administration of cisplatin. Remarkably, the intraperitoneal route showed reduced toxicity as well as improved efficacy compared with intravenous administration. Two apparent pharmacokinetic problems appear to limit the effectiveness of intraperitoneal drug administration: tissue penetration and exposure of the serosal surface. We continue our interest in the mechanisms of absorption of drugs from the peritoneal cavity and their quantitative characterization. (3) We have completed an examination the pharmacokinetics of carboplatin during and following high-flow recirculated hyperthermic perfusion of the peritoneal cavity in patients with minimal residual ovarian cancer, and the results are being published. Data on platinum concentrations in plasma filtrates and perfusate allowed calculation of the amount of drug absorbed, the permeability-area product for the peritoneal surfaces under these conditions, the total body clearance, and the pharmacokinetic advantage. Further, systemic exposure was correlated with toxicity. (4) Lung cancer originates in the epithelial cells lining the airways. It appears that there is a lengthy process leading to the development of clinically observable disease. Early detection by molecular diagnostics may provide an opportunity to intervene in the progression of the disease if appropriate agents are available and can be delivered to the epithelial cells. The regional nature of the disease suggests the possibility of chemoprevention by the administration of appropriate aerosols to the lung. Pharmacokinetic issues include the spatial distribution of agents within the lung, clearance mechanisms from the lung, and the pharmacokinetic advantage achievable by this approach. Fundamental studies of retinoid uptake and metabolism in various cell lines are expected to provide some insight into the pharmacokinetic and pharmacodynamic rationale for epithelial- directed chemoprevention. Similarly, epithelial-directed chemoprevention strategies may be useful for a variety of other malignancies. (This is a continuation of Intramural Research Project ZO1-RR- 10001-29 BEI.)